Return to Previous Section

2. Alignment Tests

On 15 July 2008 (UT) I took a series of lamp images (or "lamps" for short) using the narrow slit, without the focal reducer lens, and using 1 by 2 pixel binning. Most of this set of lamps were taken with the low-dispersion grating. I began by taking lamps with the telescope pointed at the zenith. I then took sets of lamps with the telescope slewed east by 1, 2, 3 & 4 hours. I returned the telescope to zenith, took another set of lamps, then repeated the process slewing south with increments of 15 degrees until I reached a declination of -15 degrees. I then returned to zenith and repeated the process while slewing to the west to a maximum of 4 hours. I set back to zenith, and slewed north with increments of 15 degrees until I reached the pole. I ended with a final set of lamps taken at the zenith.

The motivating purpose of this set of tests was to determine how much the zeropoint of the wavelength calibration is affected by system flexure. These data also allowed me to measure the variation of the misalignment of the spatial and spectral axes with the x and y axes of the detector.

I bias- and dark-corrected the lamps as discussed in Sections 2 & 3, above. The region of the detector exposed to the lamps defines the largest spatial extent that can be properly wavelength calibrated, and thus allowed me to refine the cropping region as discussed in Section 4.2.3. For 1 by 2 binned images at low dispersion, the first 250 rows of the detector are exposed to the emission-line lamps. Figure 37 shows two fully corrected lamps both taken at zenith. The top is an Hg lamp, and the bottom is a Ne lamp.

  • Figure 37: An 8-sec Hg lamp (upper panel) and a 2-min Ne lamp (lower panel). Both were obtained a low-dispersion, with the narrow slit and 1 by 2 pixel binning. Both have been bias- and dark-corrected and cropped.

    One can see that the spectral and spatial axes are clearly misaligned with respect to the x and y axes of the detector. Further, one can use the slope of the emission lines to determine the misalignment. However, it turns out that this will need to be done independently for every exposure. To show this, I proceeded as follows: I measured the slopes of several lines in one spectrum, and determined that all lines of a given exposure have the same slope. I then compared the slope determined from one bright line in Hg spectrum 016 with that from a bright line in Ne spectrum 015. These spectra were taken consecutively, with the same telescope pointing. The slopes were identical to within the errors. I then measured the slopes of the Hg 5480 Angstrom line on all the Hg spectra, and plotted the slopes as a function of zenith angle and exposure number. Figure 38 shows the result for the plot against exposure number. Not only does the tilt of the spectral lines change dramatically, but it does not return to the initial zenith configuration after the slew to the pole. The spatial tilt of the spectrograph changes dramatically with slews to the north, and does not return to its initial value upon return to the zenith. The lesson to take from this is that wavelength calibration exposures must be taken bracketting every object spectrum that needs wavelength calibration.

  • Figure 38: The change in x-axis pixel value as a function of y-axis pixel value for the Hg 5480 Angstrom line plotted against exposure number.

    Procede to Next Subsection

    Return to Main Line-Lamp Page

    Return to Main Wavelength Calibration Page

    Return to Outline

    Updated: 2009 August 21 [pbe]